Beginners Question - What HW do I need?


I have one question that I hope to get help with. I have just a few days experience with Cantabile and plugins, and really no experience with music production in general, or vst hosting or plugins.

I am using a couple of plugins (declick and denoise) from Acon Digital together with Cantabile to remove clicks, pops and similar noise while playing my vinyl collection. From time to time also I want to record audio. I have bought an audio interface (focusrite 2pre USB) for testing and it works great so far. Have not tested to record yet.

The laptop I’ve been testing on is a beast. The plan is to buy something smaller and hopefully much much cheaper, and more quiet. Like a Intel nuc or something similar that can do the job. Ideally it will boot directly and fire up the plugins so no screen is needed.

So the question is, what do I need to run Cantabile with the plugins and win10? Latency is not an issue, but I would not like the audio to skip.

What is the minimum processor and RAM that I need, will a nuc with Celeron or i3 processor work, or do I need to go to i5 or even an i7? Om the laptop I am using today the processor is hardly loaded at all with the setup. A few percent with fs of 48kHz.

I can always buy something cheap and test and return if it’s not working, but thought that I’ll try here first.

For example: this nuc I can get quite cheap where I live:
Celeron J3455
HD: 32GB


Modern i5 processors are fairly zippy and efficient. I do not know much about the i3 (I assume you mean the current Core i3 CPU), but here is a CPU comparison article from last year that might shed some light on the matter for you:

… pay attention to product line suffix

try to avoid mobile CPU versions.

However it is represented with ratings of course
as @terrybritton mentioned

PS I am still using an old U24E Asus which is powered by Intel Core i7 2620M
But it is on the entry level performance currently :smile:

kind regards

1 Like

I’m running a (iirc) 9th gen i5 at the studio and it’s doing great. Quad core. What you need is RAM, and lots of it. 8Gb absolute minimum, 12 is better, 16 and up is a good idea,

EDIT- I just read more carefully and if that’s actually all you;l be running then maybe you don’t need to go so high. I’d still think you need 8 to be comfortable though.

1 Like

Thanks for your replies.

I have checked more carefully now and currently my CPU is loaded at around 5%-10% (task manager) depending on sample rate, 48kHz to 192kHz.

At 192kHz it seems I get dropouts from time to time so I will be using 48kHz to 96kHz to avoid that. Not very often though. Have no clue why that happens but 48kHz or 96kHz is just fine.

From what I read in of your answers together with the dropouts at 192kHz I will go for a NUC with I5 (preferably from a later generation), with 8GB or more RAM. I will cross check against the links when I have a candidate.

Thanks again terry, Fred and Alexander!

Running your interface at 192 kHz means that your CPU needs to supply it with 4 times the audio data in the same time, compared to running it at 48 kHz - so

  • it needs to do 4 times the work in the same time, processing all your plugins on 4 times the data
  • the same buffer size gets consumed in 1/4 of the time, so the CPU needs to devote more attention to keeping the buffer filled (when the buffer runs out of data, you get droputs) vs. calculating audio data

Overall, for using VST instruments live, 48 kHz or 44.1 kHz are absolutely sufficient regarding audio quality; 192 kHz is complete overkill. I would strongly debate whether 96 kHz are at all relevant for live VST usage - there may be some merit to using 96 in very delicate studio productions, but unless you’re playing for bats or dolphins, stick with 48 or 44.1 kHz and take a load off your CPU.

To explain: a higher sample rate is necessary to improve the reproduction of very high frequencies. The Nyquist theorem states that you need to sample at twice the frequency of the highest frequency you want to reproduce correctly.

So, for 48 kHz, this means (brutally simplified) that frequencies up to 24 kHz can be reproduced correctly. Given that the hearing range of the average healthy young person ends at around 20 kHz, this should be sufficient (that’s the reason why CDs are produced at 44.1 kHz).

Going higher than that can make sense when a lot of processing stages are applied to the original audio material (typical during the production process), which may introduce artifacts that can add up to something audible - but this is for complex, multi-tracked productions with tons of plugins. I doubt anyone will notice any difference in sound between 192 and 48 in a live situation.

So do yourself and your CPU a favor and forget about 192 kHz or 96 kHz - pick one of 48 or 44.1 kHz (some older plugins seem to favor 44.1; a lot of sample material is in 44.1) and things should be groovy.




^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Agree strongly with this…